Not very polished… In Cursor, it barely worked at all even! Couldn't even call basic tools reliably, started screwing up files at turn two as if it was at the end of its context.
AGENTS
-
@testingcatalog — 2026-02-18
By
–
MyClaw agent provides execution through cross-app and system control, plus workflow automation for end-to-end tasks, running workflows across apps, files, browser flows, and system operations within your dedicated environment. Test it out for yourself
-
AI Agents as Clients: Distributed Consumption Strategy
By
–
I've published an episode on @ivoox
: "#1071: When AI agents become the client (Sell 8 Times More with Distributed Consumption) #podcast https://
go.ivoox.com/rf/167828137?u
tm_source=twitter&utm_medium=social&utm_content=audio&utm_campaign=tw_autopublish
… -
Abacus AI Desktop: Unified AI Coding and Desktop Assistant Platform
By
–
Abacus AI Desktop brings AI coding and a desktop assistant into one place. From a CLI and coding agent to chat, voice assistance, and a VS Code extension, everything works together in a single setup.
-
@testingcatalog — 2026-02-17
By
–
Control browser on Comet is now powered by the recently released Claude Sonnet 4.6 model on Perplexity Pro accounts.
-
AI voice agent handles restaurant reservations autonomously
By
–
The voice agent it generated isn't a glorified FAQ bot. It takes reservations. Modifies them. Handles special requests. "Party of 6, Saturday at 7, one guest has a shellfish allergy" Done. Synced with OpenTable.
-
LLM Agent Time Horizons: New Analysis on Model Capabilities
By
–
Ever since @METR_Evals
' fascinating work on LLM agent time horizons, I've wanted to see other attempts to draw conclusions from the same data. In a separate approach by Fengyuan and Jay, we too infer an exponential, but with shorter horizons for recent models (~2h vs METR's ~5h) -
Claude Sonnet 4.6 Launches on Databricks Platform
By
–
.
@Anthropic
’s Claude Sonnet 4.6 just landed on Databricks. With Sonnet 4.6 on Databricks, users benefit from Anthropic’s most capable agentic model, while gaining a single view of usage, spend, and performance – ensuring consistent permissions, guardrails, and rate limits -
Code Execution and Web Fetch Tools Now Generally Available
By
–
Additionally, code execution, web fetch, memory, programmatic tool calling, tool search, and tool use examples are now generally available. Read more:
-
Deploy an AI voice agent in 10 minutes from your URL
By
–
Enterprise tech used to take months. With the new Agent Wizard it takes ~10 minutes. Give it your URL. It scrapes your site. It digests your FAQs and policies. It deploys a functional voice agent while you grab a coffee. High-end AI is no longer just for the 1%. If you have a